Pavlovian fear conditioning has been widely used to reproduce anxiety-like states in laboratory animals and explore their neural substrates. Based largely on the results from these types of studies, a neural circuit model has been developed and widely adopted. Most often, these paradigms involve pairing brief stimuli such as lights or tones with footshock. Over the last several years, however, I have amassed evidence that when rats are trained and tested using longer duration stimuli (i.e., minutes as opposed to seconds), the current model is inadequate. On the basis of those data, I proposed an expanded model that focuses attention on the bed nucleus of the stria terminalis (BNST) and the stress-related peptide corticotropin releasing factor (CRF). Very briefly, the model posits that glutamatergic projections from the amygdala to the BNST are critically involved in sustained but not short-duration fear responses, and that neural transmission in these pathways is gated by CRF. A series of experiments are proposed which test specific predictions derived from this model. Experiments in Aim 1 evaluate the contribution to long- vs. short-duration fear of amygdala-to-BNST projections. Experiments in Aim 2 evaluate the contribution of CRF receptors to long- vs. short- duration fear. Experiments in Aim 3 evaluate the effect of CRF and sustained threat cues on glutamate release in the BNST, and the dependency of this release on the amygdala. The major techniques used are intra-cerebral drug infusion, permanent and reversible inactivation of selected brain regions, and the measurement of extracellular glutamate in different brain areas using microdialysis and high-pressure liquid chromatography (HPLC). The results from these studies will contribute to a more complete understanding of the neural bases of fear, which is a key step in developing new and more effective treatments for anxiety disorders such as PTSD and GAD. Moreover, as the BNST and CRF have been increasingly implicated in several other mental health and behavioral problems such as depression, drug craving, withdrawal-induced anxiety, and the emotional consequences of chronic pain, the data derived from these studies will likely be relevant to a broad range of clinically important phenomena.